Gharrett et al.: Phylogeographic analysis of mitochondrial DNA variation in Oncorhynchus kisutch 
529 
170°W 160°W 150° W 140° W 130°W 
Figure 1 
Sites from which coho salmon were sampled for mtDNA analysis. Squares denote samples 
used in both the preliminary and secondary analyses; circles denote collections used only for 
the secondary analysis. Locations are Hugh Smith River (1), Fish Creek, a Taku River tribu- 
tary (2), Berner’s River (3), Indian River (4), Ford Arm River (5), Crooked Creek (6), Little 
Susitna River (7), Buskin River (8), Karluk River (9), Eek River (10), Kanektok River (11), 
Delta Clearwater River, a Yukon River tributary (12), and Kamchatka River (13). 
bution of haplotypes and their genealogical relationships 
can provide information about the historic demography and 
gene flow of a species. Templeton and colleagues (e.g. Tem- 
pleton and Sing, 1993, Castelloe and Templeton, 1994; Tem- 
pleton, 1998) have developed methods to examine both the 
shape of the “gene tree” and the geographic distribution of 
haplotypes, which they term “nested clade analysis of geo- 
graphic distances.” 
The objectives of our study were to survey the geograph- 
ic distribution of mtDNA variation in Alaskan coho salm- 
on populations along the Gulf of Alaska and Bering Sea 
and to use that information and the mtDNA haplotype 
“gene tree” to deduce the nature of the historic demo- 
graphic processes that influenced the contemporary geo- 
graphic distribution of coho salmon. 
Materials and methods 
Coho salmon were sampled from 12 drainages in Alaska 
and one in Asia (Fig. 1). Samples of heart tissue from each 
specimen were preserved in 95% ethanol or a solution 
of 20% dimethyl sulfoxide (DMSO) and 0.25M ethylene- 
diaminetetraacetic acid (EDTA) at pH 8, saturated with 
NaCl (Seutin et al., 1991). 
Total genomic DNA was isolated by phenol-chloroform 
extraction (Wallace, 1987) or with Puregene DNA™ iso- 
lation kits (Gentra Systems Inc., Minneapolis, MN). Se- 
quences were PCR-amplified using primers that targeted 
seven regions of the mtDNA genome in pieces that range 
from about 2115 to 2689 base pairs (bp) (Fig. 2, Table 1). 
The regions were designated ND3/ND4 (including genes 
for the NADH dehydrogenase-3 subunit and NADH de- 
hydrogenase-4L and -4 subunit genes), ND5/ND6 (includ- 
ing genes for the NADH dehydrogenase-5 and -6 sub- 
units), Cytb/D-loop (including the cytochrome b gene and 
the control region), 12S/16S (including 12S rRNA gene 
and most of the 16S rRNA gene), ND1/ND2 (including 
the NADH dehydrogenase- 1 and NADH dehydrogenase-2 
subunit genes), COI/COII (including most of the cyto- 
chrome oxidase I subunit gene and the cytochrome oxidase 
II subunit gene), and A8/COIII (including genes for the 
ATPase-8 and -6 subunits and the cytochrome oxidase III 
subunit gene). The seven mtDNA regions were amplified 
by denaturation at 94°C for 5 min, followed by 30 cycles 
of 1 min at 94°C, 1 min at 55°C, and 3 min at 72°C |0.2 
mM of each dNTP, 0.2p M of each primer, 2 mM MgCl 2 , 50 
mM KC1, and 10 mM Tris-HCl, pH 8.3 with 1 unit of Taq 
polymerase (Perkin Elmer, Norwalk, CT) in a 50-pL reac- 
tion], except that amplifications of regions A8/COIII and 
ND3/ND4 required 3 mM instead of 2 mM MgCl.,. 
Subsamples of PCR products of each mtDNA region were 
digested with each of 12 restriction enzymes. The endonu- 
cleases recognized six bases ( Ase I ), multiple six-base sites 
( Ava I, Hind II, Sty I), multiple 5 base sites (BstN I), and 
four bases (RstU I, Cfo I, Dde I, Hint I, Mho I, Msp I, Rsa 
I). Digestion reactions were carried out under conditions 
recommended by the manufacturers. The resulting frag- 
